The two goals of the proposed research are to develop and validate computer based decision support systems that assist physicians (or other health care providers) in the evaluation of four ambulatory care problems and to demonstrate that the use of these systems has a beneficial effect on the cost of the evaluation process. These two goals are based on the researchers' belief that interest in cost containment should not be restricted to hospital based care. Because of the large number of physician patient encounters in ambulatory care sites, even modest reductions in cost per encounter would result in substantial savings. If used in ambulatory care settings, decision support systems could have a large impact on the cost of health care. Aid development will be based on the methodology used in an earlier study to develop a decision aid for the complaint, diarrhea of recent onset in adults. The methodology is primarily Bayesean and has three functions: to describe the level of uncertainty as to the cause of the patient's complaint conditioned on history; to select tests to reduce uncertainty at minimal cost, and; to allow the user to influence test selection depending on his concern regarding one or more diseases as causes of the complaint. Validation of the aids will be executed by comparing the aids with experts using hypothetical and actual case histories both retrospectively and prospectively. Evaluation of the aids will proceed in two phases; developmental evaluation and impact evaluation. The former will include an ongoing process of interaction of the developers with the physician clients to detect and carry out necessary changes. The impact evaluation will be executed in both hypothetical and real practice settings and will involve comparing physician output in unaided, partially aided and totally aided situations in addition to studying the effects on cost.